AWS Lambda mindmap

AWS Lambda

Run code without provisioning or managing servers; pay only for the compute time you use

• Invocation: How a function gets called
  • Synchronous: The caller waits for the response, for example API Gateway
  • Asynchronous: Lambda queues the event and retries on error; destinations and DLQs catch the results
  • Event source mappings: Lambda polls SQS, Kinesis, DynamoDB Streams, Amazon MQ, MSK, self-managed Kafka and DocumentDB
  • Function URLs: A dedicated HTTPS endpoint for a function
• Event sources
  • Amazon API Gateway: REST, HTTP and WebSocket APIs
  • Amazon S3: Object created and deleted events
  • Amazon SQS: Queue messages in batches
  • Amazon EventBridge: Event rules and schedules
  • Amazon DynamoDB Streams: Item changes in order
  • Amazon SNS: Pub/sub notifications
• Configuration: Settings that shape each execution environment
  • Memory: 128 MB to 10,240 MB; CPU scales with memory
  • Timeout: Up to 15 minutes per invocation
  • Ephemeral storage: /tmp from 512 MB to 10,240 MB
  • Architectures: x86_64 or arm64 (Graviton)
  • Environment variables: Encrypted at rest with AWS KMS
• Scaling and concurrency: How many requests run at once
  • Reserved concurrency: Sets aside a share of account concurrency for a function
  • Provisioned concurrency: Pre-initialized environments to reduce cold starts
  • SnapStart: Faster startup from a cached snapshot (Java 11+, Python 3.12+, .NET 8+)
  • Cold starts: The init phase when a new execution environment starts
• Packaging and deployment: .zip archives or container images
  • .zip archives: Up to 50 MB zipped for direct upload, 250 MB unzipped
  • Container images: Up to 10 GB, stored in Amazon ECR
  • Layers: Shared libraries and dependencies
  • Versions and aliases: Immutable versions; aliases for weighted traffic shifting
• Security
  • Execution role: What the function may do
  • Resource-based policy: Who may invoke the function
  • VPC access: Reach private resources in your VPC
  • Code signing: Deploy only trusted code
• Observability
  • Amazon CloudWatch: Logs and metrics such as duration, errors and throttles
  • AWS X-Ray: Distributed tracing